The long-term objective of the proposed studies is, through the use of biologically detailed computer simulations, to elucidate the cerebellar mechanisms of motor learning with sufficient detail to permit computer emulation of the key functional properties of this brain structure. Since evidence indicates that the human cerebellum also mediates certain cognitive functions, such information may be useful in identifying the neural basis of cognition and its malfunction in mental illness. The specific goal of the proposed studies is to use simulations to identify the information processing mechanisms employed by the cerebellum to mediate well-characterized forms of motor learning. A number of factors make this approach especially powerful: 1) the behavioral properties of these forms of motor learning are extensively characterized; 2) the synaptic organization and physiology of the cerebellum (the wiring diagram) is well known; and, 3) the way in which the simple forms of motor learning engage the cerebellum permit the input and output properties of the cerebellum to be studied, and modeled, relatively directly. With these advantages, it is possible to use biologically detailed computer simulations to address specific hypotheses regarding the cerebellar mechanisms that produce the target behaviors. Successful accounts for target behaviors by the simulations represent a level of understanding of the underlying mechanisms that is both quantitative and detailed. In contrast, identifying the limitations of the simulations should lead to crisper ideas, new theories, and should point to the experiments critical for further progress. The specific aims will test a series of detailed hypotheses regarding the mechanisms of cerebellar motor learning that are suggested by previous empirical and simulations studies. These hypotheses address how the cerebellum can learn, retain, and unlearn movements, as well as how the appropriate timing of the movements is obtained. Since the synaptic anatomy of the cerebellum is uniform across both motor and non-motor areas, identifying the cerebellar information processing involved in motor learning will also provide inroads into the information processing involved in cognition, which may lead to a better understanding of the pathologies underlying mental illness.